Eur J Clin Microbiol Infect Dis DOI 10.1007/s10096-015-2569-5
ORIGINAL ARTICLE
A survey to identify barriers of implementing an antibiotic checklist F. V. van Daalen 1 & S. E. Geerlings 1 & J. M. Prins 1 & M. E. J. L. Hulscher 2
Received: 13 October 2015 / Accepted: 28 December 2015 # The Author(s) 2016. This article is published with open access at Springerlink.com
Abstract A checklist is an effective implementation tool, but addressing barriers that might impact on the effectiveness of its use is crucial. In this paper, we explore barriers to the uptake of an antibiotic checklist that aims to improve antibiotic use in daily hospital care. We performed an online questionnaire survey among medical specialists and residents with various professional backgrounds from nine Dutch hospitals. The questionnaire consisted of 23 statements on anticipated barriers hindering the uptake of the checklist. Furthermore, it gave the possibility to add comments. We included 219 completed questionnaires (122 medical specialists and 97 residents) in our descriptive analysis. The top six anticipated barriers included: (1) lack of expectation of improvement of antibiotic use, (2) lack of expected patients’ satisfaction by checklist use, (3) lack of feasibility of the checklist, (4) negative previous experiences with other checklists, (5) the complexity of the antibiotic checklist and (6) lack of nurses’ expectation of checklist use. Remarkably, 553 comments were made, mostly (436) about the content of the checklist. These insights can be used to improve the specific content of the checklist and to develop an implementation strategy that addresses the identified barriers.
* F. V. van Daalen
[email protected]
1
Department of Internal Medicine, Division of Infectious Diseases, Centre for Infection and Immunity Amsterdam, Academic Medical Centre, Room F4-106, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands
2
Scientific Institute for Quality of Healthcare, Radboud University Medical Centre, Nijmegen, The Netherlands
Background A better use of current antibiotic agents is necessary to help control antimicrobial resistance (AMR) [1]. Antibiotic stewardship programs (ASPs) are introduced to coordinate activities to measure and improve appropriate antibiotic use in daily hospital practice [2]. Recently, a set of generic quality indicators (QIs) was developed to measure appropriate antibiotic use in the treatment of bacterial infections in adult patients hospitalised at non-intensive care unit departments [3]. The next step is using these QIs to improve antibiotic use in daily practice. Checklists are effective tools to improve patient care [4–7]. For example, the use of a surgical safety checklist resulted in an absolute risk reduction in perioperative complications of 10.6 % [4], and a large and sustained reduction in rates of catheter-related bloodstream infections was demonstrated after the implementation of an infection control checklist [5]. It can, therefore, be hypothesised that the use of an antibiotic checklist consisting of generic quality indicators (Box 1) improves antibiotic use in the hospital. However, the implementation of a checklist needs to be combined with an understanding of barriers to its uptake, since it has been shown that physicians have resistances and interferences to the use of checklists [8, 9]. Without enough attention for such barriers that hinder implementation, the use of a checklist may fail, even where other studies showed striking improvements using the same checklist [9, 10]. Ideally, potential barriers hindering uptake are analysed before implementation, to influence both the type and content of the implementation activities [11]. In general, implementation can be complicated by barriers concerning the innovation itself, the individual professional, professional interactions, the patient, the organisation, incentives, resources or the socio-political context [12, 13]. Although addressing barriers that influence the effectiveness of an innovation to
Eur J Clin Microbiol Infect Dis
improve patient care is a crucial step in the knowledge-toaction process [14], very few barrier studies have been done prior to checklist implementation [15]. Box 1: The antibiotic checklist based on generic quality indicators 1 Take at least two sets of blood cultures before starting systemic antibiotic therapy. 2 Take specimens for culture from suspected sites of infection, if possible before starting systemic antibiotic therapy, but at the latest after 24 hours of treatment. 3 Prescribe systemic antibiotic treatment according to the local antibiotic guideline. 4 a. Determine renal function. b. Adapt dose and dosing interval of systemic antibiotics to renal function if necessary. 5 Document the antibiotic treatment in the case notes or electronic medical record (EMR), including: - Indication; - Name; - Dose; - Interval; - Route of administration. 6 Determine whether antibiotic therapy can be adapted as soon as culture results become available. 7 Switch from intravenous to oral antibiotic therapy after 48–72 hours on the basis of the clinical condition, provided that oral treatment is adequate.a a
Adequate means: 1: When the antibiotic is available orally; 2: When oral intake and gastrointestinal absorption are adequate; 3: Adequate in terms of diagnosis (exceptions are e.g. endocarditis, meningitis). Barrier studies performed during or after checklist implementation [16–24] described barriers such as lack of understanding the purpose of the checklist [17], duplication with current work [18, 19] and problems with the method of implementation [24]. The present study aims to identify barriers to the uptake of an antibiotic checklist in Dutch hospitals prior to checklist implementation and to select implementation activities to target the predominant barriers that obstruct checklist uptake.
Methods We performed an online questionnaire survey among medical specialists and residents to explore anticipated barriers hindering the uptake of an antibiotic checklist.
Development of the antibiotic checklist barrier questionnaire We based our questionnaire on the Dutch validated measurement instrument for determinants of innovations (MIDI), combined with barriers found in the literature. The MIDI is developed by the Netherlands Organisation for Applied Scientific Research (TNO) and is meant as a tool for researchers to survey determinants that influence the uptake of an innovation [13]. Additionally, we performed a literature search to find publications on barriers to checklist implementation and to appropriate antibiotic use. Box 2 shows the terms we used in our search, which resulted in 168 hits in total. We selected one book [25], three systematic reviews [12, 26, 27], eight relevant barrier studies [15–22, 24] and four studies on barriers to the appropriate use of antibiotics [28–31]. Based on this information, we adapted the MIDI to fit the topic of antibiotic use, so we removed potential barriers that were not relevant and added barriers that were mentioned in the literature. Box 2: Terms for literature search Topic Reviews on barriers
Search terms in title (checklist* OR guideline*) AND (determinant* OR barrier* OR factor*) Filter: systematic reviews Barriers to checklist (barrier* OR facilitator* OR determinant* implementation OR challenge*) AND checklist* Barriers to appropriate (antibiotic* OR antimicrobial* OR antibiotic use antibacterial*) AND (barrier* OR behaviour* OR attitude*) AND (appropriate* [Title/Abstract] OR guideline* [Title/Abstract])
Hits 130
13 25
The final online questionnaire started with a description of the antibiotic checklist (Box 1) and was followed by 23 statements on anticipated barriers related to the checklist (seven items), the individual professional (six items), professional interactions (seven items), the patient (two items) and to resources (one item) (see Table 1 for the specific statements). To diminish the influence of the physician’s criticisms on the content of the checklist, statements 8 through 23 started with the sentence “Assuming that the checklist is adapted to your comments on its contents”. The level of agreement or disagreement with the statements was measured by a six-point Likert scale (1 = ‘totally agree’ and 6 = ‘totally disagree’). For each statement, it was possible to choose a seventh option: ‘I don’t know’. Furthermore, physicians could criticise the separate components of the checklist by adding comments, and there was also space for general or organisational comments. The questionnaire was completed anonymously, but we asked for the participant’s function, department and hospital.
Eur J Clin Microbiol Infect Dis Table 1 Survey questionnaire and results per domain
Na
Yes, this is a barrier (%)
This checklist explains clearly what I have to do and in which order This checklist is based on evidence or experts’ consensus
219 192
3.7 7.8
This checklist includes every step of appropriate antibiotic use in the hospital This checklist is too complex for use in daily practice
216
15.3
218
17.4
This checklist fits in current practices
217
10.1
The benefits of using the checklist are clear This checklist is feasible for all my patients who receive IV antibiotics
217 216
13.4 20.8
216
13.4
Domain
Top five
Checklist
Individual professional This checklist is a threat to my professional autonomy I expect that this checklist will improve the quality of my antibiotic prescriptions It is part of my job to use this checklist I am capable of using this checklist
212
26.9
215 211
16.7 5.2
I have enough knowledge and expertise to use the checklist adequately I have good previous experiences with working with a checklist
218 201
1.4 19.9
183
8.7
Professional interactions Colleagues will support me to use this checklist Supervisors will support me to use this checklist
177
6.8
Nurses will support me to use this checklist Colleagues will use this checklist Colleagues will expect me to use the checklist Supervisors will expect me to use the checklist Nurses will expect me to use the checklist
186 191 196 179 182
9.1 14.7 10.2 9.5 14.8
170 212
21.8 12.3
Patients Patient will be satisfied that this checklist is being used I expect that this checklist will improve the patient’s antibiotic treatment Resources There are enough financial resources to use the checklist as it is meant to be used a
☒
☒
☒
☒
☒
108 of 219 (49.3 %) answered ‘I don’t know’ → exclusion
N Number of answers after exclusion of the answers ‘I don’t know’
Setting and participants
Analysis
To gain insight into anticipated barriers to the uptake of an antibiotic checklist in hospitals prior to checklist implementation, physicians in nine Dutch hospitals were invited to participate in the survey. These nine hospitals, including two university and seven non-university hospitals, previously agreed to participate in a cluster-randomised trial on the implementation of the antibiotic checklist [32]. We visited the hospitals to inform the local antibiotic stewardship team about the antibiotic checklist and the questionnaire. Following this visit, we emailed the contact physician a link to the questionnaire, and he/she forwarded this email to the target group. The target group consisted of specialists and residents—with all levels of experience and various professional backgrounds—who have direct contact with and prescribe antibiotics to adult patients.
We included questionnaires in the analysis if at least half of the statements were appraised. We excluded statements from further analyses if ≥30 % of the participants answered ‘I don’t know’. While taking into account whether the statement was formulated as a barrier hindering uptake (“This checklist is a threat to my professional autonomy”) or as a facilitator helping uptake (“I expect that this checklist will improve the quality of my antibiotic prescriptions”), all answers (1 through 6) were re-coded into dichotomous scores: anticipated barrier ‘yes’ or ‘no’. The answers ‘I don’t know’ were excluded from the analyses. We computed frequencies and percentages and created a top five of the statements that were most often mentioned as barriers. We categorised the comments on the checklist added by the participants. If comparable comments were mentioned three
Eur J Clin Microbiol Infect Dis
times or more, the comment was considered to be relevant. We also created a top five of comments.
Results Participants The online questionnaire was filled out by participants in eight of the nine hospitals that initially agreed to participate in the cluster-randomised trial on the implementation of the antibiotic checklist. One non-university hospital no longer wanted to participate and was replaced by a similar hospital. In another hospital, the link to the questionnaire was only emailed to physicians of the department of infectious diseases. In total, 250 physicians participated in the survey, of which 219 participants completed 50 % or more of the questionnaire statements. These 219 questionnaires were included in the analyses. The participants’ characteristics are summarised in Table 2. The number of completed questionnaires per hospital ranged from 8 to 90. Barriers Table 1 shows the survey results. The statement concerning the availability of sufficient financial resources to use the checklist was excluded from further analysis, as more than 30 % of the participants answered ‘I don’t know’. The top five anticipated barriers were: (1) lack of expected quality improvement of the physician’s antibiotic prescribing (26.9 %), (2) lack of expected patients’ satisfaction with checklist use Table 2
Participants’ characteristics (n = 219) n
University/non-university Specialists/residents Specialties Internal medicine, gastroenterology and pulmonology General surgery Neurology Emergency department Urology Gynaecology Plastic surgery Oral and maxillofacial surgery Ear, nose and throat Anaesthesia Microbiology Ophthalmology Orthopaedic surgery
104/115 122/97 125 27 23 15 9 5 3 3 2 2 2 1 1
(21.8 %), (3) lack of feasibility of the checklist (20.8 %), (4) negative previous experiences with other checklists (19.9 %) and (5) the complexity of the antibiotic checklist (17.4 %). To exclude the possibility that the single hospital in which 90 physicians completed the questionnaire influenced the results disproportionally, we compared the appraisals of the 219 participants (nine hospitals) with the appraisals of 129 participants (eight hospitals). The top five anticipated barriers from these eight hospitals differed on one statement: instead of ‘complexity of the checklist’, the statement ‘nurses will expect me to use the checklist’ was in the top five. For this reason, this barrier was added to the list of frequently mentioned barriers (6). This top six contains barriers from four different domains, namely the individual professional (1 and 4), the patient (2), the checklist (3 and 5) and professional interactions (6). Comments In total 553, comments were given, of which 436 were comments and suggestions regarding the content of the checklist, 59 were general comments and 58 were organisational comments. These organisational comments described contextual factors that should be taken into account in these specific hospitals, i.e. implementation of a new electronic medical record (EMR) system (14.6 %) and merger of the hospital with another hospital (7.3 %). Comments on the content of the checklist or general comments that were relevant for all hospitals and were mentioned at least three times are presented in Table 3. The five most frequently mentioned comments were: (1c) “the item documentation leads to duplication of work” (11.0 %), (2c) “doubts about the need of blood cultures for several diagnoses” (10.5 %), (3c) “incomplete or too simplistic clarification of ‘adequate in terms of diagnosis’ for the item IV o oral switch” (8.2 %), (4c) “add information about the renal function” (6.8 %) and (5c) “add that one should take different sites for taking two different blood cultures” (6.4 %). Again, we compared the overall top five with the top five after exclusion of the 90 questionnaires of the single (university) hospital, which showed that the top five of the eight hospitals was equal to the overall top five comments. Addressing identified barriers We developed an implementation strategy that could be applied in all hospitals to address the top six anticipated barriers. The barrier (1) “lack of expected quality improvement of the physician’s antibiotic prescribing” can be addressed in the following two ways: first by showing the room for improvement, i.e. giving feedback on their current antibiotic use based on a baseline measurement, and second by providing evidence for a reduction in the length of hospital stay for the patient with adequate antibiotic use [33]. This information can be
Eur J Clin Microbiol Infect Dis Table 3
Comments per checklist item Nb
Top five
Doubts about the need of blood cultures for several diagnoses (e.g. cellulitis)
23
☒
Add in the checklist that one should take different sites for the two cultures
14
☒
Logistically difficult because of lack of time
10
Checklist item
Comment mentioned at least three times
Blood cultures (n = 88)a
Culture of suspected site of infection (n = 43)a
Prescribing antibiotics according to the hospital guidelines (n = 68)a
Doubts about the cost-effectiveness
10
Blood cultures should only be taken if the patient has fever
8
In which situation should you take more than two blood cultures?
8
Add in the checklist how long the period should be between the two cultures
6
This causes delay in the start of treatment in patients with a suspicion of bacterial meningitis Make clear that one set exists of an aerobic and an anaerobic bottle
4
Only if possible
12
The mentioned timeframe for taking the culture (
